Good morning, it's a real pleasure to be with you today and thank you Irfan for the introduction. Indeed, what I want to show you is some work that we are doing on how retail investors are allocating assets to cryptocurrencies and compare it to their investments to stocks. But before I show you the extra detail [unclear word] just to take a step back what we really are interested in this analysis is to understand the drivers of demand and with it of course the drivers of value for this new asset class, if you want, cryptocurrency. So if you think about it right, technically in asset pricing, when we think about how to value an asset, we actually try to capture the expected present values of the cash flows of any other service flows that come from the security. And then a big fraction often is of course the beliefs and the present value of changes and beliefs about the future of this asset. What I want to argue in this talk, and show you in the data is that, we know that in terms of current cash flows, cryptocurrencies and the tokens that Bitcoin or Ethereum give you as zero, so it is all about understanding what is the service flow and how do investors update about that service flow. You know, I don't want to delve too deep into this. I'm happy to do this in the Q&A but a lot of recent research for example has shown that at least the typical cryptocurrencies that are currently in circulation like Bitcoin, ETH, even say Ripple or Cardano are still actually too inefficient from a research perspective but also from availability and number of transactions per unit of time to compete with any high speed payment systems say like your credit card system, Visa and Master right. So really it's the value has to lie in either, whether people will adopt it as a store value or speculative asset. In particular, I think there are more and more people willing to believe that at least currently, cryptocurrencies are very much used as a speculative asset. especially by retail investors. And so the idea for today is to see what can be learnt about how these retail investors update about prices from their actual trades. And you know this is basically my last point here that there's obviously this famous phrase, FOMO for fear of missing out, being part of this new investment trend is something where retail investors are really chasing behind and I'll show you in our data what we can see about this.
So what I'll show you is an analysis that I have done with some of my colleagues and co-authors. Using data from a large international discount broker but it's a retail investor platform that allows trading in equities, fixed income commodities but also in cryptocurrencies. In actually many different types of cryptocurrencies but the largest ones that are being chosen and traded by investors on the platform are basically Bitcoin, ETH and Ripple. In addition, I should tell you that this is an international platform that has in particular, clients in the US and Europe. The data I will show you is from 2015 to 2020. So not the current Covid crisis, is not part of my data but as you are aware is even in the 2015 to 2020 time period there was a lot of price appreciation, price depreciation for some of the largest cryptocurrencies. And that will give me a lot of variation to actually analyse those trends. The market participants here are retail traders but there's a lot of heterogeneity. Not just in how much money people are allocating to cryptocurrencies versus other investments but also the frequency with which people trade their holding periods. So there are some people that look much more like day trader and others that are holding for a longer time period. The other thing that we see is that there is a lot of burst of activity and inactivity for retail traders which is not uncommon because we know that retail traders often have periods where they pay a lot of attention to their retail trading account and then there are times where they don't at all. The other interesting thing that will actually come in handy when I show you some of the results is that even for this stock investment or commodities and other more traditional asset classes, we actually see that people specifically focus on a few securities that they trade in. They might circle in and out of many different ones but they typically even for say in their investment don't invest in hundreds of different stocks, but they seem to focus on a few. With this let me actually give you a sense of the type of investor that we are talking about This is really a typical retail trader. This is clearly not... this is not people investing in their 401k or retirement savings account. So you see the typical investor here in an average month makes about 79 trades. Of course there are people who make much more which is at the 90th percent I bet but this is the average. But they hoard on average about 14 unique instruments, this is across stocks, crypto commodities what have you. You see that this is normal I guess, you know that the majority of the unique instruments that they hold isn't stock and then the average person invests in about two different cryptocurrencies. And the typical account, eight here at the time we meet them so to speak is about a year and a half. Now, this is just to show you as of we've seen before right, this is the concentration in the stocks we have, we see a lot of trading activity. I just put the largest 70 here for you but you can see that the top 20 stocks command most of the trading activity on this platform. This is just kind of because now what I want to show you is how investors update about prices and make allocation decisions across these different asset classes. And the headline result before I show you any details which at least we found very interesting and somewhat surprising is that investors actually are very behave very differently in their typical stock trading compared to their crypto trading. In particular, investors follow contrarian strategies in their structuring but they have a very strong momentum affinity or they display strong momentum trading in their crypto investments. So the way I will show you this and what we will looking at is the change in the allocation of wealth to either stocks or cryptocurrencies as function of past returns of the investors in my in my universe. What we'll do is I will show you this change in the allocation as the overall share. Meaning that including the appreciation of the asset. So this is basically the passive and the active allocation towards a particular security. And then I will also break out for you just the act of re-balancing. So the act of re-balancing is basically what an investor allocates in a given period to a specific asset say a specific stock or a specific cryptocurrency. But taking out without the price of appreciation so this is basically saying what is then net new money that is actively allocated to say Bitcoin as a fraction of the person's wealth. Alright, and of course this differentiation is useful because it tells us about you know, how much do people just write out a bubble on the upside or the downside that happens to them and how much do they actively make decisions about putting money into one type of asset versus another. So then with this definition what we do here, we form investor cohorts by investor type meaning as I will show you, by age, by gender, by prior experience but also by splits of whether you are a predominantly crypto investor versus stock investor or both and so on. And the reason that we chose to aggregate the data at the cohort level, meaning the time that a particular set of investors start investing in a given, I mean... with my platform here is because I wanted to get rid of any idiosyncratic variation that just comes from the fact that these retail investors sometimes become inactive and then become active again for reasons that have nothing to do with the market, but actually as I was saying before for reasons of retail investors often being inattentive but this obviously is not meaningful variation that I want to pick up on in my so this allows me smooth out idiosyncratic flows. But I can also tell you even like say I run everything at a stock level and... at a investor level the results are actually unchanged It's just more elegant if I show it to you. With this, let me show you some of the main results. The first think I want to show you is exactly this asymmetry, if you want, between the strong momentum tendency in crypto and the contrarian, the more contrarian strategy for stock. So on the left hand side of my panel here that says crypto is exactly the same analysis of how do people choose to allocate assets allocate their wealth to cryptocurrencies. But as a function of the past returns of this particular cryptocurrency and remember these are all different cryptocurrencies but the predominant ones, you really think of them as investment in Bitcoin, Ripple and ETH. And what do you see, if I show you here in the first column, this is the entire sample of crypto investments. What you see is a very strong momentum tendency. This is basically like a positive relationship between your allocation to or an investors allocation to cryptocurrencies as a function of this period return and then kind of basically cumulative return lag one month, one week, one month, three months and six months and you see all these coefficient are positive the bi-dialects or the early elects are positive but not significant. But the in particular the same period, same weeks return are very strongly momentum. What you also see which I think is quite interesting is that when I split my sample into periods with positive versus negative returns you see that this momentum behaviour of my retail investors in the crypto space is they are for up periods of strong price appreciation but also for periods of price depreciation so it's symmetric and it's very strong. Now compare the same thing here on the right side, four people's stock allocation. Here you see a very different picture so what you basically see are negative and, for the one week lag even significant results. This basically means that there is a slight contrarian strategy in how investors behave in their stock allocation. And then you also see, again I'm breaking it out into periods when the price up versus periods when the price goes down. It's in particular in periods when the price is going down when now very significantly these retail investors are, if you want, leaning against right etched in contrarian or in other words, you could say is when the price has gone up they tend to take some of the profits of take. And this is actually very much, the stock results are in line with some of the earlier work people have done in academia, using this kinds of trading broker data. This is in particular why it's actually so stunning that there is this big difference between cryptos and stock. Kind of now, remember this is the overall change. What I now want to show you is, how does this look when i just strip out the active re-balancing. Here you can really see what's going on. Again on the left is crypto, on the right are stocks. What do you see, you know among for the crypto investments is basically that in the crypto investment, my retail investors tend to not engage much in re-balancing when the price went up. If anything, what you can see is when the price went up, they tend to allocate even more of their wealth to cryptocurrencies. You know Bitcoin, Ripple and so on. While on the stock side, if you go to the right, you see exactly the opposite. You do see this active re-balancing and you see basically people taking actively wealth out of a particular stock once the price of this stock has gone up and then reallocating it to other asset class. In some sense, it seems that when you look at the crypto results here, it seems that people actually on the, when the stock is going down, this is the third column here, they are willing to just stay put but they are neither putting more money in nor take money out. However, when the price is going up, they're not just riding in out the bubble and whatever you want to call it, like the price movement. In fact they are allocating even more into the cryptocurrency so this is really reinforcing this momentum behaviour. Now, in addition, you may say maybe this is a phenomenon is this phenomenon driven by our stocks, is it driven by some of the you know maybe it's not there in some of the largest stocks that these investors invest in. And the answer is here, I'm just showing you the top 50 stocks that have been traded on this platform. You see that this behaviour is there even for these much larger stocks. So this is not a fluke somehow that on the stock side. This is driven by a few small investors. The final thing I want to show you, though to me this is really kind of very interesting is that this type of asymmetry in the crypto investment versus the stock investment is actually not an asset level phenomenon. But a person level phenomenon. What do I mean with this? You could have thought that, well look the reason you're finding this asymmetry is because the type of people who trade in cryptos are just these crazy enthusiasts and they're just very different from the people trade in stocks. The answer to this is that it's not true because what I can show you in my data and this is my last slide here is that actually when I take the sample of people who invest in both cryptos and stocks which actually is the majority of people in my sample, about 70% of the people in my sample. Now I focus basically on think of it like I take a given person if you want and I look at the crypto investment and I look at their stock investment. You still see the same strong asymmetry. You see that even for a given person, that person, they invested cryptos in a very momentum, which a very momentum strategy but will have a contrarian strategy in their stock investment. That's basically tells us that the model investor is willing to assume a very different price appreciation and also a very different price path for cryptos than for regular stocks.
So to sum up, as I showed you these retail investors follow a momentum strategy that cryptocurrency investment but that are more contrarian in their regular stock investment. This is even true when we look at people who invest in both. We also actually surprisingly found that actually that there's not much heterogeneity in this difference across people so say people with financial literacy, with more wealth, with more experience in the market still show this difference the same way as people with less experience, less money and so on. So with this, I want to end and I want to hand it back to Irfan for your Q&A. Thank you very much for your attention.
Thank you again to Antoinette for a wonderful and insightful session about the behaviours that we can learn about the crypto market from analysing data on a retail platform. Before we dive into the Q&A from Slido and as a reminder please do put your questions into the Q&A box if you do have any. Maybe if we can start with a couple selfishly. Maybe if I start with a first one, Antoinette many refer to Bitcoin as digital gold and perhaps by extension, Ethereum is digital silver. Is this a fair assessment as far as inflation hedges are there any other aspects to take into consideration with this view that has become very popular as of late?
Yeah, and this is very interesting question. I would say, I would caution investors to make this analogy. I mean to be honest even gold itself is not such a fantastic hatch as we know. But I would caution people even more when they want to make this analogy for cryptocurrency. Why do I say this? Because there are two reasons, broad reason, one I would say in the short run or the medium run if you think about cryptocurrencies the defining feature right now in their returns. One of them is that there is a lot of volatility in returns and because you have this very high level of volatility, actually it's not a very efficient hedge for anything. Because you have this massive exposure basically to price movement which makes it actually a very expensive hedge. Then I would say that's a number one thing I would have for people to have in mind. They are, in our modern advance financial market, they are much more efficient of hedging inflation risk from tips to inflation in XBOND and the lack right. But then in the long run I would also say that it's really not clear yet how much the crypto community will actually be disciplined about being an inflation hedge in the future. Because there are still a lot of governance with you know, we don't know, once Bitcoin reaches the famous 21 million coins being coined. They might decide to keep minting a few more or they might be a fork etc. Therefore, I would be a little bit cautious about this analogy.
Okay, that makes complete sense. Investment caution I think is the underlying message there. The second question I had was we noted that there were... appears to be a big retail drivers as far as price action is concerned, the digital assets as a whole. Are there any things that we can take from retail trading activity that might lend itself as an important indicator to institution investors that are looking to allocate?
Yeah that's a really interesting question. It is true I would say that retail investors seemed to have been very much a …. about it for coins, for you know kind of sussing out certain developments in the crypto space where institutions typically tend to be more cautious and therefore, slower to come in. In particular, we actually have also seen that retail investors react very strongly to changes in the regulatory and basically environment and the whole infrastructure of cryptocurrencies. So in that sense, they can actually be a useful, almost like a canary in the coal mine if you want, for even for institution investors. But I also want to say that we saw early on in crypto trading across many different cryptocurrencies exchanges that there were massive arbitrageur opportunities that were often open for many hours, even days. These were times when there were very few institution investors, particularly very few arbitrageurs in the market. It basically shows you that you need smart money, if you want, to get market... closer to [unclear word] closer to efficiency. I believe that the more we have these type of arbitrageurs in the market, these easy low hanging fruits or kind of easy returns from arbitrage have obviously been enclosed.
Nice and I think you might have coined a new crypto mean with crypto cam so maybe we look out for that in the future. And in the main body of your presentation, you spoke about proof of work cryptocurrencies in the main. Are there differences between the characteristic or the fundamental behaviours that differ for proof of stake cryptocurrencies and for our audiences' benefit can you briefly explain what the differences are between the two types of assets, and unpack the observations you might have in terms of those difference?
Yeah this is a very good point. Indeed for cryptocurrencies, at least that kind of traded on this platform. I was telling you as of now, a proof of work and a ripple effect of cause of permission of block chain. Now Ethereum number two, 2.0, will be a proof of stake. So what's the difference? Proof of work by it's name means that what the miners who are the validators of transactions and of blocks on the block chain. What they do is that they solve, if you want, a puzzle or a little math problem to get the right to validate and to confirm a block and in return they get basically, at least right now, fees plus tokens for it. What's important to understand is that in a proof of work environment entry as to become a miner is completely free. Anyone who has, you or me or anyone, who has processing capacity could become a miner and would then compete to process the blocks that are coming up on the Bitcoin block chain for example. And this is actually what makes proof of work so resource intensive because it's basically all these resources devoted to compete, to be the first one, and of course the more you bring to there the higher your likelihood that you get to be the one who mines the block. That's what makes it energy inefficient. The other thing that's important to understand is of course that a proof of work platform relies on the fact that there will always be enough even though these miners are decentralised and even actually anonymous. That there is enough decentralisation of the miner capacities so that no miner or group of colluding miners gets to control 51% of the hedging power. Because if anyone had more than 51% they could then potentially tamper with the past block chain and the new blocks being minted. Therefore, could in a way threaten the integrity of this entire block chain. In the proof of work environment, the questions are always, are we at risk of being close to a 51% attack, if you want. Then of course there is the big question of the scalability and the efficiency. This is where proof of stake comes in because proof of stake is trying to solve the issue of the resource inefficiency by saying, look. Can we get rid of or solve this problem that there are all of these decentralised minors racing to be the first solve the plot. How do we solve this in a proof of stake is to say that the only minors - or they like to call it validators, right, for proof of stake are validators who have a large stake in this particular cryptocurrency. Large stake meaning either they hold a lot of money in, say, Etherium, once it becomes a proof of stake or any other proof of stake cryptocurrency, and sometimes people also don't just look at how much did you hold right now, but how long has this entity been holding it. And here, the idea is that if hold a lot of coins in a particular cryptocurrency, you have an interest in keeping it secure and functioning. And therefore, the people who have to make the stakes should be trusted most in doing the validation, right? Well, here, you see, it becomes much more resource-efficient, but then, of course, there is a game theory puzzle to solve, which is that if some of the notes have other intentions than just maximising long-term finances, they might still get corrupted, or they may still benefit from playing a non-honest game. These are some of the dangers that also exist in a proof of stake environment. And in terms of if you look at returns in trading, I think that what we have seen is that the statistical property of the returns are relatively similar. I would say that the one thing to watch out for is that the realisation of the risk that one of these cryptocurrencies actually can be tempered with through, say, 51% a tag is slightly easier to detect in a proof of work than in a proof of stake environment. But, we have not seen this in action at this point. This is always a problem. The tier risk of the thing that you haven't seen yet manifests itself in return data sometimes only after tests happened, unfortunately.
Yeah, and that is a great point. We don't know what we don't know about the cryptocurrencies and the asset class as a whole. We do have to remind ourselves that it's relatively young, and therefore the information is still being gathered in order to be able to make those assessments. I also really enjoyed the fact that you mentioned Etherium, I'm sure John Nash and his disciples have some certain views in which cryptocurrency is developing. Without much further ado, we will now try and proceed with the Q&A. If you do have any more, please do feel free to input them and we will try to get through them as quickly as possible.
A volley of questions have come through as part of this session, so let me get cracking and get on the way as soon as possible. The first one that has come in, and it seems like it is the most popular is, "In stock market, people trade based on mean reversion, momentum, derivatives, et cetera, which requires the liquid lending market and short selling. In crypto, we do not see the liquid lending market or short selling or derivatives, how could you control those different features in your data analysis?"
Great points. So, in these retail trades, where they can control those, because I can see what people would, even on the stocks side, at least those kind of relatively unsophisticated investors are not somehow hedging other types of risk when investing, say, in Tesla, or in Apple, or you know the other stocks that they invest in. As a general fact, it is true that the infrastructure of lending market and derivatives on the crypto side are only developing, and it is also still the case that what we have seen, if you think about arbitrage in cryptocurrencies, hedging the price risks for the arbitrageurs is actually, in some sense, very expensive if they use futures, for example. And what we have seen, actually is that there is a very large private crypto lending market that arbitrageurs are accessing, and they basically tend to borrow directly cryptos from the long-term holders, the holders of the cryptocurrency, who are willing to be present-sensitive, but then, of course, share the gains from doing arbitrage trade across exchanges with dedicated arbitrageurs.
Very interesting. And maybe if I kind of dog tail one of the next questions that I am going to ask into one of the previous answers that you had around proof of work and proof of stake, we just heard about ESG earlier in one of the sessions, and the fact that cryptocurrencies use a large amount of computing power. Does crypto have …. portfolios? And how can investors reconcile with the amount of resources and drain onto the power grid that crypto uses?
So, I would say, my personal opinion that proof of work cryptocurrencies, in particular, say, like Bitcoin, does not have any other application or does not hedge any other fundamental risk in society. To me, it seems that the resource cost outstripped its benefit. I wouldn't say the same thing, say, for stable coins or smart contracts that might actually help us deliver some financial services much more efficiently. But, for the very simplistic type of cryptocurrencies like Bitcoin, I would say that the resource drain does not justify what it is potentially buying, because the benefits are very ephemeral, and our typical payment system is doing a pretty good job in a very efficient way. I also want to say, you know, just to finish this, I have seen that there are lots of "bite" people, in quotation marks, floating around somehow saying that the fact that cryptocurrencies are using so much energy is somehow a good thing, because it will incentivise the development of clean technology. Honestly, only people that don't understand economics can say this type of thing, because when you have so much additional demand, right, that is, of course, a resource drain. It is not the case that we all, as a society on this Earth are not understanding that there is massive demand for clean energy. We don't need the crypto demand on top of it. It is just a technological problem right now. They are not developing those carbon neutral or better technologies quickly enough.
Okay. That makes complete sense. We might have time for one very short question and answer here, Antoinette. So, let me choose carefully. And of course, no crypto session would be complete without a regulation question, so what impacts do you see more stringent in regulation of the cryptocurrency market having on retail participation?
Very good question. I would say that that is really the important decision of the next few months and years, right. I would say the following, and you see it playing out right now in the industry, as well. Actually, the people who understand that ultimately, wide skill adoption of any new financial technology has to be supported by robust financial regulation. Not strangulation, and obviously we also want to make sure that there is useful, constructive competition that happens in financial market. I would even point to the fact that the advent of cryptocurrencies have put pressure on the regular payment systems to drop their fees and to become, in a way, much more interested in adopting more efficient technologies. So, these are all good things. I also want us to be cognisant of the fact that there are good reasons why we have regulation in financial markets, right, such as helping consumers negative a space where there can be lots of conflict of interest. Many retail investors, even institution investors may have an information disadvantage relative to the intermediaries, and for all of this, we need regulation. And finally, there is a need, for example, regulation making sure that we won't have massive money laundering either of illegal transactions or tax avoidance behaviour. So, these are things where society would lose a lot if we didn't provide reasonable regulation of that space.
Yeah, I completely agree. In short, guard rails are important. Knowing the rules helps fundamentally build efficient and effective markets in the future. That is the end of the Q&A, I have to thank you all for the questions, there were a flurry that came through. So, apologies if we weren't able to get though everything. I'd like to thank Antoinette for her superb insights, and don't forget to leave your feedback at the end of the session. I will hand the stage back to Dwyfor.